Abstract Adoptively transferred T cells expressing chimeric antigen receptors (CARs) have shown success in eliminating hematopoietic cancers, but have so far been challenging in the solid tumor setting. A related challenge has been the identification of targets that will not be lost upon selective pressures exerted by the treatment (i.e. antigen loss variants). In this project, we propose to use protein engineering of the antigen binding domains of an antibody against novel targets that will optimize the therapy yet not be subject to antigen loss. The targets are cancer cell surface glycoproteins that contain a single, O-linked GalNAc-linkage, called a Tn antigen. These Tn-antigens are not found on the surface of normal cells, but have long been known to be associated with cancer growth and metastasis. Over the past few decades, three particular targets of Tn-linked expression, MUC1, MUC4, and MUC16 (CA125) have been described, especially as antigens on various solid tumors including ovarian cancers. In this project, we will use yeast display to engineer the Tn-dependent antibody called 237 that we have already shown can be formatted as CAR T cells, yet which can be improved for both potency and range of targetable cancers. The project will involve multidisciplinary expertise, ranging from protein engineering to cancer immunology. Guided by the known crystal structure of the 237 Fab:Tn- antigen complex, we will use yeast display and directed evolution to: i) generate 237 scFv with a range of affinities, and ii) broaden the cancer-specific reactivity to additional Tn-peptide backbones, including MUC1 and MUC4. Our Specific Aims are: Aim 1. To engineer a range of affinities and novel specificities using 237 scFv fragments as a scaffold. Aim 2. To determine the pre-clinical efficacy of engineered anti-GalNAc CARs in eradicating solid tumors. The results will provide a blueprint for the effective use and development of next- generation Tn-dependent CARs. Through this proposal, these CARs will be engineered to exhibit a high degree of specificity and activity for cancers with aberrant glycosylation due to cancer-specific mutation of Cosmc or dysregulation of glyco-transferases.